Viewpoint – On Diplomacy in Games

Interactive dialogue is a powerful system with an excellent scope, but from a development perspective, building one is a costly endeavour.

On the train to Hangzhou

Interactive dialogue is a system with which we’re all reasonably acquainted. It is a powerful system with an excellent scope that has become a frequent occurrence in games, from indies to triple-A alike. It is also a system with myriad flaws.

From a gameplay perspective, it’s captivating, from a development perspective, it’s costly. Its wide scope relies heavily on the writers to realise to its full potential.

In my opinion, one of the challenges limiting the proliferation of dialogue, or more particularly diplomacy, heavy titles is an issue of systems and design. Primarily, it is one of “action space”.


In games, the immediate goal of the player is to get to an objective using the mechanics provided by the game. Over the course of this journey, the player will encounter many smaller objectives.


Let’s create an example. This one may be unfair, and I’ll get to that later, but it highlights the point I wish to make.

Let’s imagine a door and through that door: our McGuffin. Our objective is fairly simple, to go through the door and collect the McGuffin, but there’s a small hitch in our plan: a lone soldier guards the door. For a brief moment, let’s not worry about the mechanics and just focus on the narrative.

We have a few ways to approach this problem. We could shoot this guard from a distance, engage them in epic hand-to-hand combat, or throw a magical fireball in their face. Or we could distract them with a rock we just happened to be carrying, or crawl through the vents and drop into the room like a ghost.

The actual mechanics will surely limit our imaginative exploits, and that’s not a bad thing. In many cases, the actual mechanics involved depend on and help drive the mood of the world. Resident Evil wouldn’t be in the horror category if it didn’t involve hard-to-kill enemies and limited ammunition.

Obviously, we also could have tried talking to the guard. Maybe they’re open for trading coins to let us walk on by, or maybe we can find dirt on them and leverage our way past them.

Now, the issue here is that implementing a combat or stealth system isn’t too much work, or rather, its cost is spread out between all the users of the system. A dialogue system, on the other hand, needs a considerable amount of rebuilding between uses.


The way I tend to rationalise this is that systems like combat rely on a vast state space to create engaging gameplay. The challenge with this system isn’t that it’s hard to solve; it’s that its reactive to the other sides action.


  • Let’s say we have 10 actions available to use, and the enemy 8.
  • When we can make an action, we create ten states.
  • Then the enemy would act, creating a state from each of the ten states we could have made.
  • Then we can react, adding another ten states for each state

This short back-and-forth creates 800 possible states. It doesn’t take long for that state count to “explode”. When I referred to “action space” earlier, this phenomenon is what I meant.
Even creating 800 states by hand is an arduous task. It’s also one that likely detracts from other areas of the game by way of eating available writer time.

Common Issues
Dialogue systems often run into the issue of affecting story rather than being part of moment-to-moment gameplay. To be clear, it being that way isn’t wrong; it’s just not a gameplay system.

A good example of dialogue systems used in this way is in Deus Ex. Not every mook in Dues Ex can be negotiated with, but the game makes good use of the dialogue system when it uses it. This sparing use of dialogue is a completely valid approach and one that lends itself well to the often-story-bending powers of talking your enemies down.

Another common issue in dialogue systems is a lack of cost. Deus Ex sidesteps this issue by making it occur at key story beats by design, so they don’t need a cost. For other games where you can try to negotiate with everyone, a cost is needed, lest it becomes a free Jedi mind-control. How cost is worked into this system is a tricky issue and one that needs addressing on a case-by-case basis.

The last issue I’ll touch on is one of skipping content. There is a tendency to allow skipping content based on having enough points in a stat to click a dialogue option. Removing content from your players is objectively less fun than just letting them play it.

Building Systems
One of the key similarities between systems addressed here is the idea of removing an entity from play. In stealth, an entity is removed from play when out of range. In combat, an entity is removed when incapacitated or killed.

It follows then that our system needs a way to remove an entity from it. This removal state is commonly called a “failure state” when the player lands in it. For stealth, the “failure state” often just results in combat ensuing, but it would still be a failure from a stealth perspective.

At this point, I want to define the term “immediate goal” to be the goal that an entity is trying to achieve at some core level. Additionally, the completion of this goal must result in the arrival at a “failure state” for the opposing side.

The first thing to notice here is the use of “opposing side”. For these systems to do their thing, they want to be adversarial.

For combat, both sides generally hold symmetric goals: to kill the other side.
For stealth, both sides hold asymmetric goals: one to hide, one to seek.
In both cases for these systems, the completion of the goals causes the removal of the opposite side. One of the reasons this construction works so well is that it creates conflict and resolution.

So, what do we need to build a -dialogue- diplomacy system. The change in terms here isn’t on a whim; it’s because dialogue requires written work and a “dialogue system” is commonly used to refer to the system that handles talking between characters.

Our system is not like that. Our system is about fighting with words. And it has to be about that; otherwise, it can’t create the conflict that we are after. In real life, diplomats and negotiators can strike deals that are beneficial to both sides. In games, however, that’s harder without curated dialogue.

Let’s go back to our soldier at the door. I mentioned there was a problem with it, so let's discuss that. Let’s say this guard is a hard-line, by-the-book soldier; they ain’t talking to nobody. Does this hardass just get to choose to avoid using that system at all?

This dichotomy brings the system into conflict with narrative, at least potentially. That's partially due to any conversation being the interaction of at least two people: a combat system doesn't require consent from the other side — neither does stealth — but diplomatic relations do. A hard-nosed character that wouldn't talk to the player character risks breaking immersion rather than furthering it if they break character to enable gameplay.

None of the above is to say that diplomacy systems can't be done. They can; they just need care and planning. For a game where diplomacy is a core concept, hard-nosed characters present a different type of obstacle: one where the player is required to find an alternate route rather than rely on the core mechanic.